I’m thinking that CGI could get (or has gotten?) close enough to imitating real-life appearances of humans that movies could be made in theory without any actual actors on screen. Is this impossible?: instead of Tom Cruise, or a stuntman, leaping across a chasm, Cruise simply gives permission for his face and body and voice to be duped, and an entire action film is made with CGI characters and settings that are indistinguishable (and cheaper?) than the same story filmed with real people. Presumably, if it’s cheaper, that leaves money for Cruise’s salary, and the public that can’t tell the difference visually or aurally between people/sets and CGI people/sets couldn’t give a fig how the film was made. So no one–filmgoers, actors, producers–has anything to complain about.
I’m sure they could, but there will always be people who want to see real human actors. I could see “Acted by real human actors” become a real selling point by the year 2040 or so. And I think there will always be a certain pathos element lacking. It’s not quite as moving to see something you know is animation rather than a real human portraying something.
But what if no one can tell the difference between CGI actors and real actors? What if a film’s gimmick is that Tom Cruise films a few scenes for real, and the film challenges people to tell which ones those are?
That’s the question on the table at the moment. The studios don’t object to it. The Screen Actors Guild however have drawn a line in the sand. But the technology isn’t there yet. And if the studios don’t budge, then the actors won’t come back to work.
So in the short-term this isn’t the future of movies. In the long term: the technology may get to the point where many in the audience won’t care if a talking-head occasionally slips into the uncanny valley. But that isn’t something I personally would pay money to see. And I doubt that it will get to the point where this is would be cost-effective for the studios in my lifetime.
I mean, its a gimmick. It might work once. But it isn’t something you would stake a billion dollar industry on.
Every industry everywhere seems to be moving towards trying to make money without having to deal with people, who have always been just an inconvenient obstacle to achieving extreme wealth.
They already do it with dead ones. And once they have Cruise’s body scanned, do they still need him? That’s one of the issues the actors are thinking about I believe.
It sounds plausible when you think of super popular actors that are already well established in the public mind. What happens when they age out, though? How do you establish new actors in the public mind to the point where they accept the transition to total CGI?
Further than that, however, is the fact that an actor’s popularity depends on a lot more than just screen appearance. They are full fledged human beings in the eyes of fans. They have marriages, infidelity, divorces, scandals, etc. In the end, a CGI figure is just a super sophisticated cartoon character in comparison.
I think it is a virtual certainty that AI will, at some point, be able to recreate a photorealistic performance, from scratch. Complete with dialogue and all the nuances that come from a flesh and blood person.
When that will be? Hard to say. Perhaps another couple of decades for it to truly climb out of the uncanny valley.
I think what will take longer is the eradication of humans from other parts of the process. The creation of the story that the AI characters will tell, the quality control that steers the performances and the decisions taken on what constitutes a suitable final product.
But ultimately with something this transformative that has shown exponential growth of sophistication it is hard to know quite what is coming and when.
This is just me, but all of that can be easily done with CGI, Photoshop, deepfakes….how would you know if it’s ever really happening? I know it sounds paranoid.
Well, probably because there is an entire industry dedicated to telling the public anything and everything that goes on with production companies, directors, actors, their off screen lives, etc. I can’t possibly imagine something that significant being kept a secret from the public. If nothing else, someone would “drop the dime” for a fistful of cash.
I can definitely foresee it partially replacing actors for reshoots, added scenes, alternate endings—all the stuff that currently requires them to get all the actors back on a plane to wherever, after they thought they were done, to address whatever panic the test audiences gave the producers.
For a while facial images, facial expressions, and most body motion in general are best accomplished by using live models. It is much more practical to scan a human head or body, and to capture motion of live actors than to consume many man hours by artists and animator accomplishing a lesser quality result. However, after an initial modeling session the images and movements can be used and reused, modified and remodified repeatedly, and whole new characters can be created out of combinations of recordings of different people. In Deep Fake now we can see one person’s facial expressions used to modify those of another.
If you’re in the motion picture or television business you may still keep using unique and creative actors who deliver the same kinds of audience response we see now. If your in the online service business you don’t need to have your animated people expressing the deep loss of a loved one, or moments of great physical agony, The next advertising character like Progressive’s Flo could well be someone who spent a few days modeling in a studio for all the imagery used in 1000 commercials.
Purely generated speech is still not that great either. A voice actor could spend a day or two recording enough words, sounds, and intonations of them to recombine them into speech that can’t easily be detected as computer generated, but in movies and TV actual voices may still be preferred.
It is difficult because of a concept called The Uncanny Valley Effect. That is people’s uncomfortable feeling when seeing animated humanoids that aren’t quite human. We can detect the tiny bits of unnatural motion and sound both consciously and unconsciously and it leads to uncomfortable feelings.
For both audio and video the idea of a programmable actor based on generic qualities and then refined by a director may sound like a fantastic technological approach to movie making but it may always be more cost and time efficient to just record humans acting like humans for a desired result. That is until AI is good enough to do without directors, not something right around the corner like the rest of it.
And is this so far from the actors perhaps showing up for a few days, trying on a few costumes, getting photographed from various angles, and perhaps giving a few line readings, getting paid, and then “BYE, thanks, that’s all we need”? Instead of a few months shooting, it’s now a few days, or a few hours, for the same pay. Some directors dream of eliminating all the hassle of dealing with actors on a set or on location.
It will start with animation and work it’s way up to blockbusters. And once you have the technology to generate movies there’s no need to pay licensing to Cruise or Clooney. Just develop a face and personality that people like and push it. The studios have been doing that since the beginning.
And it’s trivial to internationalize: language, nationalities, locations, etc.
I think you will still get movies with human actors and are advertised as such. They might style them more simply, more like plays, to differentiate them.
We’ve made such progress, technologically, in CGI over the past two decades that it’s impossible for me to imagine that we won’t be advanced immeasurably further in ten or twenty years. The trend towards literal comic books being the source of movies over that time span leads me even more certainly to the conclusion that actors will be unnecessary to make movies that appeal to young people eager for action-packed entertainment that doesn’t really require subtle performances from gifted actors.